DeepSeek Plans $1.5 Billion Fundraise and Prepares for IPO
Taylor Wilson
DeepSeek is negotiating a $1.5 billion round at a $71 billion valuation and plans to IPO as early as late this year — barely a month after closing its first outside funding, a pace that is resetting how the market prices Chinese AI.
Two rounds in one month — why is the money coming this fast?
DeepSeek closed its first external round at roughly $7 billion, valuing it above $50 billion, around late May.
If this round lands at $71 billion, the valuation jumps 42% in a matter of weeks.
This means → investors are not pricing DeepSeek on a steady climb — they are racing to get in before the next markup.
Who is investing, and what does the cap table signal?
Existing backers include Tencent and the Beijing National AI Industry Investment Fund.
Tencent ties DeepSeek directly to China's largest internet traffic platform. The state fund's presence marks the company as a national strategic asset.
In plain terms = DeepSeek is collecting more than capital — it is locking in distribution, data access, and political cover in a single cap table.
How close is DeepSeek to U.S. leaders in the enterprise market?
In June, DeepSeek handled about 23% of the trillions of tokens routed through Vercel — an enterprise AI gateway that helps companies call AI models. Anthropic handled roughly 32%.
This means → DeepSeek's enterprise penetration is within 10 percentage points of a top U.S. AI lab.
This reflects a broader shift: enterprises are choosing AI on cost-performance, not brand or geography.
How is DeepSeek solving the chip problem?
Its cloud services currently run on chips made by Huawei — a critical "de-Americanization" path as U.S. export controls keep tightening.
Reuters has reported that DeepSeek is also designing its own AI chips to further reduce reliance on outside suppliers.
In plain terms = Huawei chips are the stopgap; in-house silicon is the long-term independence play. Running both tracks signals DeepSeek is preparing for the worst-case supply scenario.
It rejected venture capital for years — what changed?
Founded in 2023, DeepSeek rattled Silicon Valley with its V3 and R1 models, matching top U.S. model performance at a fraction of the cost.
Per The Information, the company resisted outside funding until April this year, when Anthropic's technical breakthrough shifted its calculus.
This means → DeepSeek did not raise because it was short of cash — it raised because the competition accelerated. The IPO is about stockpiling ammunition for the next leg of the AI arms race.
What is the IPO timeline, and what is still uncertain?
The target window is as early as late 2025, no later than 2027, but the exact timing remains open.
Whether the current round closes on schedule is itself unconfirmed. DeepSeek has not commented on the reports.
In plain terms = both the fundraise and the IPO are still "planned," not "done" — but the sheer pace is already repricing how the market values Chinese AI companies.
Content is for reference only, not financial advice.